Objectives: Atrial fibrillation (AF) is associated with hemostatic abnormality and increased risk of thromboembolic events. The aim of the present study was to evaluate whether elevated D-dimer levels in combination with clinical risk factors can adequately predict subsequent thromboembolic events in patients with AF. Method: This was a post hoc analysis of the single-center, prospective observational study. Consecutive patients with nonvalvular AF (245 patients, 74 ± 9 years) treated with warfarin were included. D-dimer levels were measured in combination with clinical risk factors to assess the relationship of this parameter with subsequent thromboembolic events. Results:D-dimer levels were elevated (≧0.5 µg/ml) in 55 (22%) patients. During an average follow-up time of 756 ± 223 days, 9 (1.8%/year) thromboembolic events occurred. Cox proportional hazard analysis showed that elevated D-dimer levels (hazard ratio, HR, 14.3, p < 0.01) as well as high CHADS2 score (≧3; HR 9.43, p < 0.01) was associated with thromboembolic events. C-statistic improved to 0.848 from 0.781 when D-dimer level was added to the predictor model based on CHADS2 score. Elevated D-dimer levels also predicted thromboembolic events in the subgroup of patients with high CHADS2 score (n = 76, HR 12.1, p = 0.021). Conclusion: CHADS2 scores and D-dimer level can be a useful marker of subsequent thromboembolic events in patients with AF during oral anticoagulant therapy and combination of these can effectively predict thromboembolic events.